Mailing List Archive

Mythepg's death as a result of mythfilldatabase failures
Hey guys!

One of the problems that reliably happens for me is the failure of
mythfilldatabase to update all of the channel info (~600 chans on
satellite!!).. So, sometimes a single channel's entries are missing, and
sometimes more..

Anyway, I *suspect* this is the reason why mythepg keeps segfaulting.. Any
ideas on how I can confirm this?

tarek : )
Re: Mythepg's death as a result of mythfilldatabase failures [ In reply to ]
f its just single programs missing then i doubt it
i often have individual programs missing
On Fri, Oct 25, 2002 at 11:19:52AM -0400, tarek Lubani wrote:
> Hey guys!
>
> One of the problems that reliably happens for me is the failure of
> mythfilldatabase to update all of the channel info (~600 chans on
> satellite!!).. So, sometimes a single channel's entries are missing, and
> sometimes more..
>
> Anyway, I *suspect* this is the reason why mythepg keeps segfaulting.. Any
> ideas on how I can confirm this?
>
> tarek : )
>
> _______________________________________________
> mythtv-dev mailing list
> mythtv-dev@snowman.net
> http://www.snowman.net/mailman/listinfo/mythtv-dev
Re: Mythepg's death as a result of mythfilldatabase failures [ In reply to ]
I have been playing around, and it is almost certainly not something
missing in the database.. I've checked through it semi-carefully..

The seg-faults are consistent and very invasive.. They happen when i go
from 100 to 865 backwards (100 is my first channel and 865 is my last)..
And also at many times throughout, such as whenever 204 (CNNHN) shows up..
If, for example, I go to 205, then do mythepg, there is no segfault.. what
else can cause mythepg to segfault? Here is a backtrace. This is just as
relaible as the others, and happened when I went from channel 100 to 865 by
hitting the up cursor once.:

[root@courage orangey]# gdb mythepg
GNU gdb 5.2.1-2mdk (Mandrake Linux)
This GDB was configured as "i586-mandrake-linux-gnu"...
(gdb) run
Starting program: /usr/local/bin/mythepg
3 57 669
Program received signal SIGSEGV, Segmentation fault.
0x4003ac70 in GuideGrid::paintPrograms(QPainter*) () from /usr/lib/libmyth.so.0
(gdb) bt
#0 0x4003ac70 in GuideGrid::paintPrograms(QPainter*) ()
from /usr/lib/libmyth.so.0
#1 0x40038cdf in GuideGrid::paintEvent(QPaintEvent*) ()
from /usr/lib/libmyth.so.0
#2 0x402db420 in QWidget::event(QEvent*) ()
from /usr/lib/qt3/lib/libqt-mt.so.3
#3 0x40243c19 in QApplication::internalNotify(QObject*, QEvent*) ()
from /usr/lib/qt3/lib/libqt-mt.so.3
#4 0x402434c1 in QApplication::notify(QObject*, QEvent*) ()
from /usr/lib/qt3/lib/libqt-mt.so.3
#5 0x4021565c in QWidget::repaint(QRegion const&, bool) ()
from /usr/lib/qt3/lib/libqt-mt.so.3
#6 0x402445c9 in QApplication::sendPostedEvents(QObject*, int) ()
from /usr/lib/qt3/lib/libqt-mt.so.3
#7 0x40244488 in QApplication::sendPostedEvents() ()
from /usr/lib/qt3/lib/libqt-mt.so.3
#8 0x401da875 in QApplication::processNextEvent(bool) ()
from /usr/lib/qt3/lib/libqt-mt.so.3
#9 0x40245270 in QApplication::enter_loop() ()
from /usr/lib/qt3/lib/libqt-mt.so.3
#10 0x40427aa3 in QDialog::exec() () from /usr/lib/qt3/lib/libqt-mt.so.3
#11 0x08049669 in main ()
#12 0x408e7082 in __libc_start_main () from /lib/i686/libc.so.6


>f its just single programs missing then i doubt it
>i often have individual programs missing
Re: Mythepg's death as a result of mythfilldatabase failures [ In reply to ]
On Friday 25 October 2002 01:24 pm, tarek Lubani wrote:
> I have been playing around, and it is almost certainly not something
> missing in the database.. I've checked through it semi-carefully..
>
> The seg-faults are consistent and very invasive.. They happen when i go
> from 100 to 865 backwards (100 is my first channel and 865 is my last)..
> And also at many times throughout, such as whenever 204 (CNNHN) shows up..
> If, for example, I go to 205, then do mythepg, there is no segfault.. what
> else can cause mythepg to segfault? Here is a backtrace. This is just as
> relaible as the others, and happened when I went from channel 100 to 865 by
> hitting the up cursor once.:

Please recompile everything for debugging (swap the comment on the top two
lines of settings.pro), and regenerate the backtrace. What you sent is
pretty much useless..

Isaac